Upload
others
View
4
Download
0
Embed Size (px)
Citation preview
Basic Residues R260 and K357 Affect the ConformationalDynamics of the Major Facilitator Superfamily MultidrugTransporter LmrPWei Wang, Hendrik W. van Veen*
Department of Pharmacology, University of Cambridge, Cambridge, United Kingdom
Abstract
Secondary-active multidrug transporters can confer resistance on cells to pharmaceuticals by mediating their extrusionaway from intracellular targets via substrate/H+(Na+) antiport. While the interactions of catalytic carboxylates in thesetransporters with coupling ions and substrates (drugs) have been studied in some detail, the functional importance of basicresidues has received much less attention. The only two basic residues R260 and K357 in transmembrane helices in theMajor Facilitator Superfamily transporter LmrP from Lactococcus lactis are present on the outer surface of the protein, wherethey are exposed to the phospholipid head group region of the outer leaflet (R260) and inner leaflet (K357) of thecytoplasmic membrane. Although our observations on the proton-motive force dependence and kinetics of substratetransport, and substrate-dependent proton transport demonstrate that K357A and R260A mutants are affected in ethidium-proton and benzalkonium-proton antiport compared to wildtype LmrP, our findings suggest that R260 and K357 are notdirectly involved in the binding of substrates or the translocation of protons. Secondary-active multidrug transporters arethought to operate by a mechanism in which binding sites for substrates are alternately exposed to each face of themembrane. Disulfide crosslinking experiments were performed with a double cysteine mutant of LmrP that reports thesubstrate-stimulated transition from the outward-facing state to the inward-facing state with high substrate-binding affinity.In the experiments, the R260A and K357A mutations were found to influence the dynamics of these major proteinconformations in the transport cycle, potentially by removing the interactions of R260 and K357 with phospholipids and/orother residues in LmrP. The R260A and K357A mutations therefore modify the maximum rate at which the transport cyclecan operate and, as the transitions between conformational states are differently affected by components of the proton-motive force, the mutations also influence the energetics of transport.
Citation: Wang W, van Veen HW (2012) Basic Residues R260 and K357 Affect the Conformational Dynamics of the Major Facilitator Superfamily MultidrugTransporter LmrP. PLoS ONE 7(6): e38715. doi:10.1371/journal.pone.0038715
Editor: Eugene A. Permyakov, Russian Academy of Sciences, Institute for Biological Instrumentation, Russian Federation
Received March 9, 2012; Accepted May 11, 2012; Published June 20, 2012
Copyright: � 2012 Wang and van Veen. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permitsunrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: H.W.v.V. is grateful for funding by the Biotechnology and Biological Sciences Research Council (BBSRC). The funders had no role in study design, datacollection and analysis, decision to publish, or preparation of the manuscript.
Competing Interests: Corresponding Author Hendrik W. van Veen is a PLoS ONE Editorial Board member. This does not alter the authors’ adherence to all thePLoS ONE policies on sharing data and materials.
* E-mail: [email protected]
Introduction
Multidrug resistance originating from active efflux of pharma-
ceuticals from cells is a major obstacle in the treatment of
microbial infections and human cancers [1–4]. Active drug
extrusion is mediated by multidrug transporters, which recognize
a wide variety of structurally unrelated compounds [5–7]. On the
basis of bioenergetic and structural criteria, multidrug transporters
can be divided into two major classes: (i) primary-active ABC
transporters, which utilize the free energy of ATP binding/
hydrolysis to efflux toxic substrate, and (ii) secondary-active
transporters, which mediate drug extrusion in a coupled exchange
with H+ and/or Na+ ions [8]. There are many questions regarding
the mechanism by which these transport systems operate.
The multidrug transporter LmrP from Lactococcus lactis is a
secondary-active transporter that is a member of the widely
occurring Major Facilitator Superfamily (MFS) [9]. LmrP is a 408-
amino-acid polypeptide that is predicted to contain 12 transmem-
brane helices (TMHs). Its drug efflux activity was first found in
genetic screens based on its ability to confer resistance on
Escherichia coli to high concentrations of a monovalent amphiphilic
ethidium cation [10]. The transporter can utilise both the
membrane potential (Dy, interior negative) and the chemicalproton gradient (DpH, interior alkaline) of the proton-motive force(Dp) to mediate the efflux of amphiphilic substrates from cells. Thetransport mechanism was inferred to be electrogenic antiport,
whereby monovalent cationic ethidium+ and tetraphenylpho-
sphonium (TPP+) move in exchange for two or more protons
[11]. In the absence of the Dp, LmrP can also mediate facilitatedtransport of substrates down their chemical gradient [12,13].
During active and facilitated transport, LmrP is thought to
mediate the transbilayer movement of substrates via a rigid body
motion of the N- and C-terminal halves with a concomitant
alternating access of substrate-binding sites to the inside surface
and outside surface of the cytoplasmic membrane, in an analogous
fashion as proposed for LacY, LeuT and other secondary-active
transporters [14–17].
Studies on drug/proton antiporters including EmrE, MdfA, and
QacA have highlighted the importance of carboxyl residues in
TMHs in interactions with cationic substrates and/or protons [18–
PLoS ONE | www.plosone.org 1 June 2012 | Volume 7 | Issue 6 | e38715
20]. Previously a 3-D homology model of LmrP with high
geometrical quality was constructed [12], based on the high-
resolution, inward-facing structure of the glycerol-3P/Pi antiporter
GlpT from Escherichia coli [21]. GlpT and LmrP are evolutionaryrelated in an unambiguous manner, and the modeling suggested
that the proteins have a comparable protein fold. Similar to the
GlpT template, inward-facing LmrP is predicted to expose an
internal cavity between its N- and C-terminal halves to the inner
leaflet of the cytoplasmic membrane. The carboxylates exposed on
the surface of this cavity in LmrP are organized into two distinct
clusters containing additional neutral-polar and aromatic residues
(see Fig. 2 in [12]). Cluster 1 in the C-terminal half contains D235
and E327 in immediate proximity of each other, and is located
near the apex of the cavity, while Cluster 2 in the N-terminal half
contains D142. Functional analyses of mutants suggest that D142
acts a dedicated proton-binding site, whereas D235 and E327 are
part of a flexible multidrug binding surface that can interact with
cationic substrates and protons [22].
In contrast to the more extensively characterized roles of
carboxylates in secondary-active multidrug transporters, few studies
have focused on the role of basic residues in these systems. LmrP is
predicted to contain only two basic residues in TMHs, R260 (TMH
8) and K357 (TMH 11) in the C-terminal half (Fig. 1). Different from
the catalytic carboxylates (D142, D235 and E327) that are present on
the surface of the interior chamber in our inward-facing model, K357
and R260 are located on the outer surface near the phospholipid
head group region of the inner leaflet (K357) and outer leaflet (R260)
of the cytoplasmic membrane. Here, we characterized the functional
role of K357 and R260 in multidrug transport by LmrP.
Results
Expression of LmrP MutantsTo study the functional role of K357 and R260 in the C-
terminal half of LmrP, we generated single mutant proteins in
which these residues were each replaced by a nonpolar A. In
addition, K357 and R260 were substituted in a charge-conserva-
tive fashion (Table 1). Wildtype (Wt) LmrP and K357A, K357R,
R260A, and R260K mutant proteins were expressed in L. lactis
strain NZ9000 DlmrA DlmrCD [23] under the control of a nisin A-inducible promoter in the expression plasmid pNZ8048 [24].
Coomassie-stained SDS-PAGE and western blot analysis of equal
amounts of total membrane proteins in inside-out membrane
vesicles (ISOVs) demonstrated that the expression level of the
mutant proteins in the cytoplasmic membrane was similar to that
of Wt (Fig. 2A). As the presence of basic residues in TMHs of
membrane proteins can act as potent determinants of membrane
topology [25,26], we compared the topology of the C-terminal half
of Wt and mutant LmrP proteins through studies on the
accessibility of the intracellular C-terminal His6-tag to digestion
with proteinase K in preparations of well-defined ISOVs and
right-side-out membrane vesicles (RSOVs). Similar to Wt LmrP,
complete proteolytic removal of the His-tag from the mutant
LmrP proteins was achieved in ISOVs, but not in RSOVs
(Fig. 2B). These data demonstrate the intracellular location of the
C-terminus for all proteins. Together with the observed function-
ality of the mutants in Dp-dependent and facilitated transportmodes (see below), these results suggest that the topology of LmrP
remains intact following the substitutions of K357 and R260.
Mutations Affect Ethidium TransportEthidium is a permanent monovalent cation, and a known
substrate of LmrP. The dye undergoes a fluorescence intensity
enhancement upon its intercalation in intracellular nucleic acids.
The effects of K357 and R260 replacements on LmrP-mediated
ethidium transport were first monitored in intact cells. When ATP-
depleted L. lactis DlmrA DlmrCD cells, preloaded with 2 mMethidium, were incubated with 25 mM glucose, a Dp (interioralkaline and negative) was generated by the proton pump activity
Figure 1. Stereoview of LmrP showing positions of basic residues R260 and K357. Ribbon representation of inward-facing LmrP [12]viewed sideways from the external side of the membrane. Transmembrane helices (TMHs) 1 to 6 in the N-terminal half (in grey) and TMHs 7 to 12 inthe C-terminal half (in green) are numbered. R260 (in light blue, TMH 8) and K357 (in dark blue, TMH 11) are labelled, and are shown in stickrepresentations. The three catalytic carboxylates D142 (TMH 5), D235 (TMH 7) and E327 (TMH 10) are coloured magenta, red, and orange,respectively.doi:10.1371/journal.pone.0038715.g001
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 2 June 2012 | Volume 7 | Issue 6 | e38715
of the F1FO-H+ ATPase. The LmrP proteins couple this Dp to
ethidium extrusion from the cell. The efflux activity for K357R
mutant (2.5160.43 a.u./s) was similar to that of Wt LmrP(2.5860.63 a.u./s). Although reduced compared to Wt LmrP,the R260K mutant (0.3960.14 a.u./s), R260A mutant(0.2260.03 a.u./s) and K357A mutant (0.2860.08 a.u./s) exhib-ited a substantial rate of ethidium efflux compared to control
(Fig. 3A).
To study downhill, facilitated transport of substrates by LmrP in
the absence of Dp, the facilitated influx of ethidium in intact cellswas measured at an external ethidium concentration of 2 mM. Theinflux rates for K357R mutant (1.9260.12 a.u./s) and K357Amutant (1.9860.12 a.u./s) were similar to Wt LmrP(1.7860.30 a.u./s) (Fig. 3B). The influx rates for the R260Kmutant (0.3160.05 a.u./s) and R260A mutant (0.1460.02 a.u./s)were decreased compared to that of Wt LmrP (P,0.05) (Fig. 3B).However, all mutants mediated significantly more ethidium influx
than control (0.0460.02 a.u./s) (P,0.05) (Fig. 3B).To further analyse the functional role of K357 and R260, we
examined the ability of the mutant proteins to export ethidium in
the presence of single components of the Dp, i.e. either the DpH orthe Dy, through the addition of ionophores valinomycin ornigericin (1 mM each) [27]. Valinomycin is a mobile carrier in themembrane that catalyses the electrogenic movement of K+ and
dissipates the Dy when cells are suspended in high K+ (50 mM)buffer. In contrast, nigericin selectively dissipates the DpH underthese conditions by catalyzing electroneutral K+/H+ exchange
across the membrane. Our studies revealed that dissipation of
either component of the Dp did not significantly affect Wt LmrP-mediated transport from cells preloaded with 2 mM ethidium(Fig. 4A,B). In the case of the K357A mutant, dissipation of the
DpH by nigericin had a small effect on ethidium efflux whereasvalinomycin significantly attenuated transport activity (Fig. 4C).
This finding indicates that the K357A mutation induces a shift to
Dy dependence over DpH dependence compared to Wt LmrP.Finally, the addition of nigericin gave the most significant
reduction in ethidium efflux by R260A LmrP compared to the
addition of valinomycin (Fig. 4D). Compared to Wt LmrP, the
R260A mutation therefore induces a shift to DpH dependenceover Dy dependence in ethidium efflux by the mutant. The
energetics of ethidium efflux by the K357R mutant was similar to
Wt LmrP, whereas the energetics of the R260K mutant was
similar to that of the R260A mutant (data not shown).
Mutations Affect Hoechst 33342 TransportSubstrate transport by the LmrP mutants was also studied in
ISOVs. Hoechst 33342 is a substrate of LmrP [28] and many
other MDR transporters [29,30]. In aqueous solution Hoechst
3342 is essentially non-fluorescent, whereas its partitioning in the
membrane is associated with a remarkable enhancement of
fluorescence [31]. The fluorescence change of this reaction is
reduced at acidic pH, which is most likely due to inhibitory effects
of protonation of Hoechst 33342 on the fluorescence quantum
yield and on membrane partitioning. In our experiment in ISOVs,
a rapid fluorescence intensity enhancement was observed follow-
ing the addition of 0.25 mM Hoechst 33342 (Fig. 5) due to thepartitioning of the dye in the membrane. The addition of ATP to
the ISOVs leads to the generation of a Dp (inside acidic andpositive) through inward proton pumping by the F1FO-H
+
ATPase, which can be used by LmrP proteins to drive proton/
Hoechst 33342 antiport. Upon the addition of ATP, ISOVs
containing Wt LmrP showed a rapid Hoechst 33342 fluorescence
quenching compared to control (Fig. 5) due LmrP-mediated
transport of dye from the membrane into the lumen of the ISOVs.
The activity of the K357R mutant was closest to that of Wt LmrP.
Although reduced, the transport activities of the K357A, R260A
and R260K mutants remained substantial (Fig. 5). These findings
were consistent with those in ethidium transport measurements
(Fig. 3). Further experiments focussed on Wt LmrP and the
K357A and R260A mutants.
Kinetics of Ethidium Transport by LmrP and MutantsUsing established methods for LmrA [32,33] and MsbA [34],
the kinetic parameters of the facilitated ethidium influx reaction
mediated the LmrP proteins were determined in ATP-depleted
Table 1. Primers used for site-directed mutagenesis.
No Primer Sequence 59-39
1 C270S Fw C ATA TAT TTG ATT TTA GCT AGC GTA TTA GTT GTC
2 C270S Re GAC A AC TAA TAC GCT AGC TAA AAT CAA ATA TAT G
3 I34C Fw CT TCA ATG ACC TGT TAT TAT AAT C
4 I34C Re G ATT ATA ATA ACA GGT CAT TGA AG
5 V240C Fw GAT AAT TTT TTA CCT TGC CAT TTA TCG AAT AG
6 V240C Re CT ATT CGA TAA ATG GCA AGG TAA AAA ATT ATC
7 K357R Fw GCA ATT AGA ATG CCA ATT GCA TCA ATT C
8 K357R Re TGG CAT TCT AAT TGC TGC AAC GCC ATT A
9 K357A Fw GGC GGT GCA GCA ATT GCA ATG CCA ATT
10 K357A Re AAT TGG CAT TGC AAT TGC TGC AAC GCC
11 R260K Fw T GGA CAA AAG ATG CTC ACC ATA TAT TTG
12 R260K Re AG CAT CTT TTG TCC ATA GAT TTC AAA AA
13 R260A Fw GAA ATC TAT GGA CAA GCG ATG CTC ACC
14 R260A Re GGT GAG CAT CGC TTG TCC ATA GAT TTC
doi:10.1371/journal.pone.0038715.t001
Figure 2. Expression and topology of mutant LmrP proteins. A,Western blot of total membrane proteins in inside-out membranevesicles (ISOVs) (15 mg of protein/lane) probed with anti-His5-tagantibody, showing equal expression of Wt and mutant LmrP proteinsin the plasma membrane of L. lactis NZ9000 DlmrA DlmrCD (with anapparent molecular mass on 10% SDS-PAGE of about 32 kDa), and lackof this expression in controls cells. B, The availability of the C-terminalHis-tag in Wt LmrP, and the K357A and R260A mutants to cleavage byproteinase K (+ PK) at the external side of ISOVs and right-side-outmembrane vesicles (RSOVs) was tested during incubations of 10, 20 and30 min. Subsequently, remaining C-terminal His-tag was detected byWestern blotting. Control incubation (lane 1, -PK) is without proteinaseK. Equal loading of the lanes was confirmed by densitometry of thesame samples on Coomassie-stained SDS-PAGE (not shown).doi:10.1371/journal.pone.0038715.g002
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 3 June 2012 | Volume 7 | Issue 6 | e38715
cells using ethidium concentrations between 0.25 mM and 10 mM(Fig. 6A). Statistical analysis of the kinetic data revealed a
significant increase in the Vmax of influx for the K357A mutant
compared to Wt LmrP (2.9760.22 a.u./s for K357A mutant vs.1.7860.18 a.u./s for Wt LmrP) (Fig. 6B) without a significantchange in the apparent Km for ethidium (1.4260.29 mM forK357A mutant vs. 0.9260.30 mM for Wt LmrP) (Fig. 6C). TheR260A mutation significantly decreased the Vmax of influx
(0.1860.03 a.u./s for R260A mutant vs. 1.7860.18 a.u./s forWt LmrP) (Fig. 6B), whereas no significant change was observed
for the apparent Km for ethidium (1.4960.68 mM for R260Amutant vs. 0.9260.30 mM for Wt LmrP) (Fig. 6C).
Facilitated efflux of ethidium in cells was measured by
imposing an outwardly directed ethidium concentration gradient
in the absence of a Dp. ATP-depleted cells in 2 ml of 50 mMKPi buffer (pH 7.0) containing 5 mM MgSO4 (A660 of 0.5) were
equilibrated with ethidium at concentrations ranging from
0.25 mM to 10 mM for 1 h at 30uC in the final volume of400 ml of the KPi buffer. After equilibration, cells were pelleted
by centrifugation. The assay was started by resuspension of the
cell pellet in 2 ml of the same buffer without ethidium, and
fluorescence was followed over time. The kinetic analysis
(Fig. 6D) showed that the Vmax of facilitated ethidium efflux
mediated by K357A mutant is comparable with that of Wt
LmrP (0.7860.22 a.u./s for K357A mutant vs. 1.0660.25 a.u./s for Wt LmrP) (Fig. 6E). In addition, the apparent Km for
ethidium was not significantly affected by the K357A mutation
(4.3762.57 mM for K357A mutant vs. 7.4963.09 mM for WtLmrP) (Fig. 6F). However, the R260A mutation significantly
decreased the Vmax of efflux compared to Wt LmrP
(0.1460.04 a.u./s for R260A mutant vs. 1.0660.25 a.u./s forWt LmrP) (Fig. 6E), whereas the apparent Km for ethidium was
unaltered (5.4463.11 mM for R260A mutant vs. 7.4963.09 mMfor Wt LmrP) (Fig. 6F). The lack of a significant change in Kmfor ethidium in the efflux reaction by the LmrP mutants
compared to Wt is consistent with the influx data (see above),
and suggests that R260 and K357 are not critical for ethidium
binding by LmrP.
Figure 3. Ethidium transport in intact cells. A, ATP-depleted cells expressing Wt LmrP, the mutants K357R, K357A, R260K, or R260A, or no LmrPprotein (control) were preloaded with 2 mM ethidium in 50 mM KPi buffer (pH 7.0) containing 5 mM MgSO4. Glucose (25 mM) was added at the timepoint indicated, to allow the generation of metabolic energy in the cells. The initial rates of active efflux were calculated over the first 20 s duringwhich transport was linear. B, Facilitated ethidium influx in ATP-depleted cells. Ethidium bromide (2 mM) was added at the time point indicated. Theinitial influx rates were calculated over the first 40 s of linear uptake with the exception of K357A for which the rate was determined over the first20 s. Data in the main text represent the mean 6 SEM of three independent experiments (n = 3). Ethidium transport was measured by fluorimetry.doi:10.1371/journal.pone.0038715.g003
Figure 4. Dependence of active ethidium efflux on the composition of the proton-motive force (Dp). Ethidium transport was measuredas described in the legend to Fig. 3A for cells without LmrP (control) or with Wt LmrP, or expressing mutant K357A (panel A) or R260A (panel B).Composition of Dp (interior negative and alkaline) in cells in KPi buffer was manipulated by the addition of 1 mM of the ionophores nigericin (+N) orvalinomycin (+V) that dissipate the DpH and Dy, respectively. Traces represent observations in at least four independent experiments (n = 4).doi:10.1371/journal.pone.0038715.g004
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 4 June 2012 | Volume 7 | Issue 6 | e38715
Proton TransportThe potential role of R260 and K357 in LmrP-mediated proton
translocation was also assessed by measuring the change in the
intracellular pH (pHin) during facilitated influx of substrate in
ATP-depleted cells loaded with the fluorescent, covalently-bound
and non-transported pH-sensitive probe 5 (and 6-)-carboxyfluor-
escein succinimidyl ester (CFSE). Although ethidium is often used
as a test substrate for LmrP, its excitation and emission spectra
overlap with those of CFSE. Therefore, the non-fluorescent cation
benzalkonium+ was used as an alternative substrate. During
facilitated influx of benzalkonium by Wt LmrP in ATP-depleted
cells at an external concentration of 44.1 mM, the fluorescence ofthe pH probe increased compared to the non-expressing control
(Fig. 7A), pointing to an increase in the intracellular pH (pHin)
during proton/benzalkonium antiport. The addition of the
benzalkonium also resulted in an increased fluorescence of CFSE
in cells expressing the K357A or R260A mutants compared to
control (Fig. 7A). Calculation of pHin values from the linear
fluorescence increase during the first 40 s of facilitated benzalk-
onium influx by Wt LmrP and K357A LmrP, and the first 150 s
for R260A LmrP and Control, show that the pHin increased from
6.4560.03 to 6.8160.02 for Wt LmrP, 6.4160.15 to 6.9560.19for mutant K357A, and 6.9960.02 to 7.2560.05 for mutantR260A, whereas the pHin in the control cells remained relatively
constant between 6.6660.11 to 6.7760.16 during the experiment(Fig. 7B). Hence, for all LmrP proteins, the pHin was significantly
elevated compared to the pHin in control cells (Wt LmrP and
R260A LmrP vs. Control, P,0.05; K357A LmrP vs. Control,P,0.01; K357A LmrP vs. R260A LmrP, P,0.05). The change inpHin in cells was largest for Wt LmrP and the K357A mutant, and
smallest for the R260A mutant (Fig. 7C), consistent with the effect
of these mutations on the Vmax of substrate transport (Fig. 6E).
Taken together, the data that K357 and R260 are not critical for
proton translocation.
Conformational ChangesTo investigate whether the changes in the Vmax of substrate
transport by the K357A and R260A LmrP mutants are related to
changes in conformational transitions between inward-facing and
outward-facing states, the conformational changes in the K357A
and R260A mutants were compared with LmrP itself in a Cys-less
background (LmrP-cl) in which cysteine residues were inserted at
Figure 5. Hoechst 33342 transport in ISOVs containing WtLmrP, LmrP mutants or no LmrP (control). ISOVs were diluted to aconcentration of 0.2 mg protein/ml in 100 mM KPi (pH 7.0) containing5 mM MgSO4, 0.1 mg/ml creatine kinase and 5 mM phosphocreatine.After 50 s, Hoechst 33342 was added to a final concentration of0.25 mM, followed by the addition of 2 mM ATP at 150 s. Hoechsttransport was measured by fluorimetry. Traces represent observationsin three independent experiments (n = 3).doi:10.1371/journal.pone.0038715.g005
Figure 6. Kinetic analysis of LmrP-mediated, facilitated ethidium influx and efflux. A, The ethidium influx rates in ATP-depleted cells weremeasured as a function of the ethidium concentration. Each data point represents the mean 6 SEM of four independent experiments (n = 4). B,Comparison of Vmax of the K357A or R260A mutant with that of Wt LmrP. The data were analysed by one-way ANOVA followed by the post-hocTukey’s test. **P,0.01, ***P,0.001, both significantly different. C, Comparison of Km for ethidium of the K357A or R260A mutant with that of WtLmrP. Analysis by one-way ANOVA shows that Km values are not significantly different. D, The efflux rates in ATP-depleted cells were measured as afunction of the ethidium concentration. Each data point represents the mean 6 SEM of three independent experiments (n = 3). E, Comparison of Vmaxof K357A or R260A mutant with that of Wt LmrP. The data were analysed by one-way ANOVA followed by the post-hoc Tukey’s test. *P,0.05,significantly different. F, Comparison of Km of the K357A or R260A mutant with that of Wt LmrP. Km values were not significantly different when thedata were analysed by one-way ANOVA.doi:10.1371/journal.pone.0038715.g006
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 5 June 2012 | Volume 7 | Issue 6 | e38715
position I34 and V240 (double cysteine mutant LmrP-cl I34C/
V240C, referred to as LmrP-dc) (Fig. 8). Recently, an analogous
strategy was incorporated in studies on conformational changes in
the MFS transporter LacY [35,36]. According to our structure
model of LmrP, the Cys side chains at positions I34 and V240
would be close enough to form a disulfide crosslink in the inward-
facing conformation (Fig. 8) but not in the outward-facing
conformation where, based on analogous outward-facing models
for secondary-active membrane transporters such as LacY, LeuT
and Mhp1 [15,16,37], these residues would be separated by at
least 20 Å. In an ethidium transport assay in which cells were
preloaded with 2 mM ethidium, the LmrP-cl mutant and theLmrP-dc mutant retained significant active ethidium efflux activity
compared to Wt LmrP (2.2260.05 a.u./s, 1.3860.05 a.u./s, and5.6360.40 a.u./s, respectively, P,0.001) (Fig. 9A). Hence, theLmrP-dc mutant can be used to study conformational changes
associated with substrate transport.
When total membrane protein from RSOVs was incubated
under oxidizing conditions by exposure to copper phenanthroline,
and subsequently separated on SDS-PAGE gel and analysed by
Westernblot probed with anti His-tag antibody, LmrP-cl protein
was detected with a single signal corresponding to a molecular
mass of about 32 kDa (Fig. 9B, lane 1). In contrast, if this
experiment was repeated with LmrP-dc, this double cysteine
mutant gave rise to two signals. Whereas the upper signal
corresponds to the same molecular mass as LmrP-cl and represents
uncrosslinked protein, the lower signals runs at a smaller apparent
molecular mass (Fig. 9B, lane 3). Interestingly, when followed over
an extended incubation period of 10 min, the intensity of the
lower signal progressively increased whereas the intensity of the
upper signal reduced during this time (Fig. 9C). The notion from
these experiments that this lower signal was associated with
intramolecular disulfide crosslinking in LmrP-dc was confirmed by
the observation that the lower band was resolved into the upper
band when the protein samples were diluted in loading buffer
containing dithiothreitol (DTT), prior to separation on the SDS-
PAGE gel (Fig. 9B, lane 5). We anticipate that the intramolecular
crosslinking in LmrP results in its incomplete unfolding in
denaturing SDS-PAGE, causing a shift to a lower apparent
molecular mass.
The presence of the transported substrates ethidium or
benzalkonium stimulated the disulfide crosslinking reaction in
LmrP-dc. Unpaired student t test for the density of upper and
lower band signals in Fig. 9B (lane 3) after crosslinking for 5 min in
the absence of substrate indicated that less than half of the
population of LmrP-dc was in the crosslinked state (crosslinked
LmrP-dc versus uncrosslinked LmrP-dc: 44.461.8% vs.55.661.8%, P,0.05) (Fig. 10A). In the presence of 97.9 mMbenzalkonium (Fig. 9B, lane 4), a substantially larger population of
LmrP-dc was found in the crosslinked conformation (crosslinked
LmrP-dc vs. uncrosslinked LmrP-dc: 62.263.4% vs. 37.863.4%,P,0.01) (Fig. 10A). In a complementary comparison, pairedstudent t tests for density of the lower band signal in the absence
and presence of benzalkonium indicated that the population of
crosslinked LmrP-dc significantly increased (44.461.8% vs.62.263.4%, P,0.05) (Fig. 10A), whereas paired student t test
Figure 7. Determination of change in CFSE fluorescence and pHin in intact cells during facilitated benzalkonium influx. A, ATP-depleted cells containing Wt LmrP, LmrP mutants (R260A or K357A) or no LmrP protein (Control) were preloaded with 30 mM CFDASE ester, which isintracellulary covalently bound as CFSE, and diluted to a final A660 value of 0.5 in 2 ml of 50 mM KPi/5 mM MgSO4 at pH 7.0. After the CFSEfluorescence measurement was started, 44.1 mM benzalkonium chloride was added at the time indicated by the arrow. Base line in grey is added tofacilitate comparison of the initial slopes (see main text). B, The pHin values in cells before and after the addition of the benzalkonium were derivedfrom fluorescence calibration curves (not shown). All values represent the mean 6 SEM of three independent experiments (n = 3). C, The changes inpHin were analysed using the one-way ANOVA followed by the post-hoc Tukey’s test. *P,0.05, **P,0.01, significantly different. All values representthe mean 6 SEM of three independent experiments (n = 3).doi:10.1371/journal.pone.0038715.g007
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 6 June 2012 | Volume 7 | Issue 6 | e38715
for densitometry of upper band signal in the absence and presence
of benzalkonium suggested that the population of uncrosslinked
LmrP-dc significantly decreased (55.661.8% vs. 37.863.4%,P,0.05) (Fig. 10A). In contrast to these findings for LmrP-dc,the exposure of LmrP-cl to benzalkonium did not have any effect
on the mobility or intensity of its signal (Fig. 9B, lane 2). Similar
observations were made with ethidium (data not shown), but the
band shifts were stronger with benzalkonium.
Consequently, the data show that at the end of the 5-min
period of crosslinking, the dominant population of LmrP-dc in
the RSOVs is in an outward-facing conformation. Upon the
addition of a transport substrate, the two six-helix bundles
Figure 8. Close proximity between I34 and V240 in inward-facing LmrP. Stereoview of inward-facing model [12] viewed from externalmembrane surface, showing I34 (TMH 1) in yellow and V240 (TMH 7) in brown, and the basic residues R260 (TMH 8) in light blue and K357 (TMH 11) indark blue. I34 and V240 were substituted by C in cysteine-less LmrP (referred to LmrP-cl), yielding the double cysteine mutant I34C V240C LmrP-cl(referred to as LmrP-dc).doi:10.1371/journal.pone.0038715.g008
Figure 9. Intramolecular disulfide formation in LmrP-dc (I34C & V240C) and mutants derived thereof. A, Ethidium efflux in intact cellsexpressing Wt LmrP, LmrP-cl or LmrP-dc show that, similar to Wt LmrP, the mutants are transport active. The measurements were performed asdescribed in the legend to Fig. 3A. The traces represent results of three independent experiments (n = 3). B, Cysteine crosslinking between I34C andV240C in LmrP-dc. RSOVs containing LmrP-dc (2 mg protein/ml) in 100 mM KPi buffer, pH 7.0, plus 5 mM MgSO4 were exposed to 0.5 mM copperphenanthroline (Cu/Ph) in the presence or absence of 97.9 mM benzalkonium chloride (BC) for 5 min at 10uC. The reactions were stopped by theaddition of excess (10 mM) of the thiol alkylator N-ethylmaleimide followed by incubation for 2 min at 10uC. Total protein (5 mg) of RSOVs from thecrosslinking reactions were mixed with 5 6SDS sample-loading buffer devoid of dithiothreitol (DTT) (lane 1–4) or containing 2 mM DTT (lane 5), andseparated on 11% SDS-PAGE. The SDS-PAGE gel was subsequently subjected to Western blot transfer onto a HybondP membrane. Transferredproteins were probed with anti-His5 antibody. Upper band: uncrosslinked protein; Lower band: crosslinked protein. C, Time course for disulfideformation between I34C and V240C. The crosslinking in RSOVs containing LmrP-dc was followed for 10 min at 10uC as described for panel B, lane 3.Samples of the crosslinking reactions (2 mg) were mixed with 5 6 SDS sample-loading buffer devoid of DTT, and separated on SDS-PAGE. D, Theeffect of K357A or R260A mutation on disulfide formation between I34C and V240C. The crosslinking was performed as described in panel B for 5 minat 10uC. Equal loading of lanes was confirmed by densitometry of the same samples on Coomassie-stained SDS-PAGE (not shown).doi:10.1371/journal.pone.0038715.g009
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 7 June 2012 | Volume 7 | Issue 6 | e38715
formed by the N- and C-terminal halves are drawn closer
together at the external side of the protein due to high-affinity
binding of this substrate to the inward-facing state, thus
increasing the population of LmrP-dc in the inward-facing
conformation. In the inward-facing state, the I34C and V240C
side chains are close enough to form an intramolecular disulfide
bond.
Effect of K357A and R260A Mutations on CrosslinkingThe cysteine crosslinking assay was used to examine the
conformational transitions of the K357A and R260A LmrP-dc
mutants and the effect of benzalkonium on these transitions. For
the experimental data for K357A LmrP-dc in the absence of
benzalkonium (Fig. 9D, lane 1), the unpaired student t-test for
densitometry of signals indicate that at the end of the 5 min period
of crosslinking, the population of crosslinked protein was smaller
than the population of the uncrosslinked protein (28.664.1% vs.71.464.1%, P,0.01) (Fig. 10B). The presence of benzalkonium(Fig. 9D, lane 2) significantly increased the population of cross-
linked K357A LmrP-dc compared to its absence (61.765.9% vs.28.664.1%, P,0.05), whereas the population of uncrosslinkedLmrP significantly decreased (71.464.1% vs. 38.365.9%,P,0.05) (Fig. 10B). Hence, as observed for LmrP-dc, benzalk-onium binding to LmrP-dc K357A in RSOVs shifts the population
of LmrP from outward-facing to inward-facing.
For the LmrP-dc R260A mutant in the absence of benzalk-
onium, the population of crosslinked protein was much larger than
that of uncrosslinked protein (82.961.3% vs. 17.161.1%,P,0.001) (Fig. 9D, lane 3; Fig. 10C, D). Strikingly, in the presence
of benzalkonium, this difference remained largely unaltered
(82.961.1% vs. 17.463.6%) (Fig. 9D, lane 4; Fig. 10C, D).Meanwhile, paired student’s t-test for the density for the
uncrosslinked protein in the absence and presence of benzalk-
onium remained almost the same (17.161.1% vs. 17.463.6%)(Fig. 10C). These observations show that by the end of the
crosslinking reaction with LmrP-dc R260A, the predominantpopulation in the RSOVs is in an inward-facing conformation,
and that this distribution is unaffected by the addition of
benzalkonium (Fig. 10C). These findings contrast with those on
K357A LmrP-dc, which in the absence of added benzalkoniumformed a significantly smaller population of crosslinked protein
compared to LmrP-dc itself (28.664.1% vs. 44.461.8%, P,0.05)(Fig. 10D), suggesting a reduced population of K357A LmrP-dc inthe inward-facing conformation. Hence, the data indicate that
substitution of K357 and R260 to A shifts roughly equal
populations of inward-facing and outward-facing LmrP in the
crosslinking reaction towards enlarged populations of outward-
facing LmrP (K357A) and inward-facing LmrP (R260A), respec-
tively.
Discussion
In our transport experiments, K357A and R260A mutations did
not affect the Km for ethidium (Fig. 6). However, the R260A
mutant showed a substantial reduction in the Vmax compared to
Wt LmrP and the K357A mutant (Fig. 6). This difference in Vmaxwas also reflected in the proton transport activities during
benzalkonium-proton exchange, which were low for the R260A
mutant compared to Wt LmrP and the K357A mutant (Fig. 7).
Figure 10. Densitometric analyses of disulfide formation by LmrP-dc, K357A LmrP-dc and R260A LmrP-dc. The white and black barsrepresent the relative signal intensity of the upper and lower bands, respectively, on the Westernblots in panels B–D in Fig. 9, relative to the sum ofupper and lower bands signals (with an intensity of 100%). A, LmrP-dc, B, K357A LmrP-dc, C, R260A LmrP-dc, in the absence or presence ofbenzalkonium chloride. D, Relative signal intensity of the lower band (reflecting inward-facing conformation) for LmrP-dc, K357A LmrP-dc and R260ALmrP-dc, respectively, in the absence of benzalkonium chloride. Error bars represent 6 SEM and are based on three independent experiments.Unpaired student t tests were used for panels A, B, and C, to compare the density of upper and lower signals under the same treatment, whereaspaired student t-tests were used to compare the density of the upper band signal or lower band signal in the absence and presence of benzalkonium.One-way ANOVA and the post-hoc Tukey’s test were performed for the analyses in panel D, to determine whether the density of the lower bands ofthe LmrP mutants were significantly different from each other. *P,0.05, ***P,0.001, significantly different. Upper band: uncrosslinked protein; Lowerband: crosslinked protein.doi:10.1371/journal.pone.0038715.g010
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 8 June 2012 | Volume 7 | Issue 6 | e38715
The findings indicate that while K357A and R260A mutations
affect ethidium-proton and benzalkonium-proton antiport, neither
K357 nor R260 is critical in substrate binding or proton
translocation. This conclusion might relate to the difference in
the location of K357 and R260 in our inward-facing LmrP model
compared to the location of the catalytic carboxylates D142, D235
and E327 with established roles in substrate binding and/or
proton coupling [22]. Whereas the catalytic carboxylates co-
localise on the surface of the interior cavity [12], K357 and R260
are present on the outside surface of LmrP, where they are
exposed to the phospholipid bilayer.
The intra-molecular cysteine crosslinking experiments with
LmrP-dc (Fig. 9) are in agreement with the current view thatmembers of the MFS can be present in at least two major
conformational states that are interconverted by the rigid-body
movement of both halves of the transporter, i.e., the inward-facing
conformation in which I34C and V240C in LmrP can interact
with each other, and the outward-facing conformation in which
these cysteines are too far apart for cross-linking (Fig. 8). Together
with the correct predictions by our LmrP model on the catalytic
role of D142, D235 and E327 in proton and substrate interactions
[12,22], the observed cross-linking between the two cysteines in
LmrP-dc (Fig. 8) further supports the accuracy of the model (Fig. 8)
[12]. The substitution of K357 and R260 by A changed the
distribution of LmrP-dc over these protein conformations during
the crosslinking reaction (Figs. 9,10). At the end of the reaction, the
population of K357A LmrP-dc in the outward-facing state was
larger than for LmrP-dc, but the transition to the inward-facing
state could still be established by the addition of benzalkonium
(Fig. 10). Under identical conditions, R260A LmrP-dc was mostly
present in the inward-facing conformation, and the presence of
benzalkonium did not induce conformational transitions
(Figs. 9,10).
As K357 and R260 are located close to the head group
region of the phospholipid bilayer, they might affect the
conformations of LmrP through direct interactions with
phospholipids in the interfacial region. In X-ray crystal
structures of various membrane proteins, such as the yeast
cytochrome c oxidase [38] and cytochrome bc1 complex [39],
protein-lipid interactions have been observed that are based on
electrostatic interactions between the phosphodiester layer of the
membrane and a basic amino acid side chain. In crystal
structures of the mammalian voltage-sensitive K+ channels
KvAP [40,41] and Kv1.2 [42], and the lipid- and mechano-
sensitive K+ channel K2P TRAAK [43], this type of interaction
was found to be critical for the stability of conformational states
that occur during gating. Our finding that the Dp-dependentethidium and Hoechst 33342 transport activity of mutants with
conservative R260K and K357R replacements was closer to Wt
LmrP than non-expressing control, whereas the activity of the
R260A and K357A mutants was closer to the non-expressing
control than Wt LmrP (Figs. 3,5) points to an important
functional role of a basic side chain at these positions. It is
noteworthy that R260 is highly conserved in orthologs of LmrP
in species including Streptococcus, Enterococcus, Staphylococcus,
Macrococcus, Bacillus, Weissella, Leuconostoc and others. Amongst
these orthologs, the conservation of K357 appears to be more
limited to those in Weissella and Leuconostoc species.
The notion that the K357 and R260 influence transitions
between the inward-facing state and outward-facing state of
LmrP could explain the energetics of active ethidium efflux by
the K357A and R260A mutants (Fig. 4). Although it is known
that LmrP can utilise the Dp (interior negative and alkaline) todrive ethidium efflux from cells, it is not established in detail
how Dy and DpH dependence are related to the conforma-tional changes during the drug transport cycle. If the
carboxylates responsible for proton coupling (D142 in TMH5,
D235 in TMH7, and E327 in TMH10 [22]) are protonated
when exposed to the outside surface of the membrane and
deprotonated in the inward-facing state, the transition from the
outward-facing state to the inward-facing state would be DpHdependent. Our recent experiments imply that this is indeed the
case [22]. By analogy to voltage-gated K+ channels [42], basic
residues in TMHs in LmrP could impose Dy dependence whenthey would move in a protonated (cationic) state from the
electrical outside to the electrical inside. However, LmrP does
not contain basic residues in central regions of the TMHs that
form reorienting surfaces, and it is therefore more likely that Dydependence in LmrP originates from the reorientation of the
dissociated (anionic) carboxylates during the transition from the
inward-facing state to the outward-facing state (Fig. 11). Our
previous observations on LmrP that any single carboxyl to
neutral amide replacement at positions D142, D235 or E327
changes electrogenic 3H+-propidium2+ antiport to electroneutral
2H+-propidium2+ antiport [22], is consistent with this proposal.
Hence, the Dy and DpH might act on different parts of thetransport cycle (the transition from 2 to 3, and the transition
from 1 to 2, respectively). The finding that the K357A mutation
diminishes the population of LmrP in the inward-facing
conformation (state 2 in Fig. 11) might hamper the Dy-dependent progression of the transport cycle from the inward-
facing state to the outward-facing state (state 3 in Fig. 11),
which could offer an explanation for the bias towards Dydependence of ethidium efflux by this mutant (Fig. 4A).
Likewise, as the R260A mutation reduces the population of
LmrP in the outward-facing conformation (state 1 in Fig. 11),
this phenotype might hinder the DpH-dependent progression oftransport cycle from the outward-facing state to the inward-
facing state, and might thus increase the DpH dependenceduring ethidium efflux (Fig. 4B). It can be anticipated that the
replacement of residues in loop regions that connect TMHs
might also affect the distribution of transport proteins over
conformational states, when these residues support the stability
of these states through loop-loop and/or loop-helix interactions.
Past conclusions on the direct role of D68 and D128 in the
intracellular loop between TMH 2 and 3, and TMH 4 and 5,
respectively, in proton-coupled transport by LmrP ([44,45])
might therefore require further investigations.
In conclusion, our data on LmrP suggest that K357 and R260
are not directly involved in interactions with transported substrates
or coupling protons. Instead, these residues affect the dynamics of
major protein conformations in the transport cycle, potentially
through their interactions with phospholipids and/or other
residues in LmrP. The replacement of these residues by alanine
therefore changes the Vmax at which the transport cycle can
operate. The K357A and R260A substitutions also influence the
energetics of transport, which might relate to the notion that
transitions between conformational states are differently affected
by components of the proton-motive force.
Materials and Methods
Construction of MutantsThe mutations K357R, K357A, R260K, or R260A were
generated in C-terminally His6-tagged LmrP in the E. coli
pGEM-5Zf(+) (Promega) derivative pGHLmrP [12] by site-directed mutagenesis using the QuikChange kit (Stratagene)
with forward and reverse primers as listed in Table 1. Using the
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 9 June 2012 | Volume 7 | Issue 6 | e38715
same method, the endogenous C270 in LmrP was replaced by S
to create the cysteine-less protein (LmrP-cl) (Table 1). The PCR
product was digested with NcoI and BamHI, and cloned into
pGH10ABCG2 (an E. coli vector encoding C-terminally His10-
tagged ABCG2) in exchange for the ABCG2 gene, yielding
pGH10LmrP-cl. The double-Cys I34C/V240C mutant of LmrP
(LmrP-dc) was generated in pGH10LmrP-cl using the primers
listed in Table 1. Finally, the K357A and R260A mutants were
generated in pGH10LmrP-dc using the primers listed in Table 1.
The cloned PCR products were sequenced to ensure that only
the intended changes were introduced. The mutant lmrP genes
were then subcloned as NcoI-XbaI fragments into pNZ8048
[24] downstream of the nisin A-inducible promoter, yielding
pNZLmrP K357R, pNZLmrP K357A, pNZLmrP R260K,
pNZLmrP R260A, pNZLmP-cl, pNZLmrP-dc, pNZLmrP-dc
K357R, pNZLmrP-dc K357A, pNZLmrP-dc R260K, and
pNZLmrP-dc R260A.
Bacterial Strains, Plasmids and Growth ConditionsL. lactis strain NZ9000 DlmrA DlmrCD [23], which is devoid of
the endogenous multidrug transporters LmrA and LmrCD, was
used as a host for pNZ8048-based plasmids [24]. Cells were grown
in M17 broth (Oxoid Ltd., Basingstoke, UK) supplemented with
0.5% glucose at 30uC. Where required, chloramphenicol wasadded to a final concentration of 5 mg/ml. Medium wasinoculated with a 1:50 dilution of an overnight cell culture, and
cells were grown to an A660 of 0.5–0.6. Expression of proteins was
then induced for 1 h at 30uC in the presence of a 1:1000 dilutionof the supernatant of an overnight culture of the nisin-producing
strain L. lactis NZ9700 by previously described methods [12]. E.
coli strain XL1 Blue was used as a host for pGHLmrP and
derivatives, and was grown at 37uC under aerobic conditions on
an orbital shaker in Luria-Bertani (LB) Broth (Difco) containing
50 mg/ml carbenicillin where required.
Preparation of Inside-out Membrane Vesicles (ISOVs)ISOVs were prepared as described previously [22], suspended
in 100 mM KPi (pH 7.0) containing 10% glycerol, and stored in
liquid nitrogen. The protein concentration was determined using
the Bio-Rad DC assay (BioRad, Richmond, CA, USA), and
expression of the His-tagged proteins was confirmed on Western
blot probed with anti-His5-tag antibody (Sigma-Aldrich, St. Louis,
MO, USA).
Preparation of Right-side-out Membrane Vesicles (RSOVs)RSOVs were prepared by osmotic lysis as described previously
[46], suspended in 50 mM KPi buffer (pH 7.0) containing 10%
glycerol at a protein concentration of about 5–10 mg/ml, frozen
in liquid nitrogen and stored at 280uC until use. The proteinconcentration was determined using the Micro BCA Protein Assay
(Thermo Scientific Pierce, Rockford, IL, USA), and expression of
the His-tagged proteins was confirmed on Western blot probed
with anti-His5-tag antibody (Sigma-Aldrich, St. Louis, MO, USA).
ImmunoblottingTotal membrane protein in ISOVs and RSOVs were subjected
to 10% SDS-PAGE. Proteins were then electroblotted to Hybond-
P membrane (Immobilon Transfer Membrane; Milipore Corp.,
Billerica, MA, USA) and were detected in the presence of a 1:1000
dilution of the monoclonal anti-His5-tag antibody (Sigma Aldrich).
The primary antibody was probed with a horseradish peroxidise
(HRP)-coupled secondary anti-mouse antibody (Sigma-Aldrich).
Detection of the secondary antibody was performed using the
enhanced chemiluminescence system (ECL; Pierce, Rockford, IL,
USA).
Protease Digestion of C-terminal Histidine-tagTo compare the accessibility to protease digestion of the
intracellular His-tag of Wt LmrP versus mutant LmrP proteins in
ISOVs and RSOVs, 95 mg total membrane protein in 47.5 ml of50 mM (K)HEPES buffer (pH 7.5) supplemented with 1 mM
CaCl2, was incubated in the presence of 3.8 mg proteinase K [27].After incubation at 37uC, the reaction was terminated at the timesindicated in Fig. 2B by the addition of 2.5 ml of 200 mMphenylmethanesulphonyl fluoride (PMSF, in ethanol), and then
mixed with 10.4 ml of 6x SDS-loading buffer. Total membraneprotein (15 mg) of each sample was subjected to SDS-PAGEfollowed by immunoblotting using the monoclonal anti-His5-tag
antibody.
Ethidium Transport in Intact CellsCells were grown to an A660 of 0.5–0.6 from overnight cultures
in fresh M17 medium at 30uC. The cells were harvested bycentrifugation at 6,500 g for 8 min at 4uC and washed once with50 mM KPi buffer (pH 7.0) containing 5 mM MgSO4. Cells were
resuspended to an A660 of 1.2 in the same KPi buffer containing
0.5 mM of the uncoupler 2,4-dinitrophenol, and then incubated
for 45 min at 30uC [32]. The cells were subsequently washed threetimes with ice-cold 50 mM KPi buffer containing 5 mM MgSO4at pH 7.0, and resuspended in this buffer to an A660 of 5.0 and
kept on ice.
To measure facilitated ethidium influx, ATP-depleted cells were
diluted 10-fold in 2 ml of 50 mM KPi buffer containing 5 mM
MgSO4 at pH 7.0. After 80 s, the required concentration of
ethidum bromide ranging from 0.25 mM to 10 mM was added to
Figure 11. Alternating access model for Dp-dependent ethi-dium efflux by LmrP. Proton extrusion by the F1FO-H
+ ATPase (inpurple) in cells generates a Dp (interior negative and alkaline) across thecytoplasmic membrane. The N- and C-terminal halves of LmrP arerepresented by yellow ellipses. Minus signs in pink represent thecatalytic carboxylates (D142, D235 and E327) that are involved insubstrate/proton interactions. During electrogenic efflux via proton-ethidium antiport, carboxylates are protonated in the outward-facingconformation (1), which triggers a DpH-dependent conformationalswitch to the inward-facing conformation (2). Drug binding and protonrelease at the inside surface will facilitate a second conformationalswitch in which dissociated carboxylates reorient back to the outward-facing conformation (3) in a Dy-dependent fashion. Conformation (3)might be the same as conformation (1). K357 (red plus sign, TMH 11)and R260 (blue plus sign, TMH 8) are located on the outer surface ofLmrP, near the phospholipid head group region of the inner leaflet(K357) and outer leaflet (R260) of the membrane. In our experiments,alanine replacements at position 357 and 260 shift roughly equalpopulations of inward-facing and outward-facing LmrP towardsoutward-facing (K357A) and inward-facing (R260A) populations.doi:10.1371/journal.pone.0038715.g011
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 10 June 2012 | Volume 7 | Issue 6 | e38715
the cells, and fluorescence was followed over time in a Perkin
Elmer LS 55B fluorimeter using excitation and emission wave-
lengths of 500 and 580 nm, respectively, and slit widths of 5 and
10 nm, respectively. The influx rates in Fig. 6A were calculated
over the first 25 s of uptake during which uptake was linear. Rates
of passive influx by nonexpressing control cells were subtracted
from those obtained for cells expressing Wt LmrP or LmrP
mutants. To measure facilitated ethidium efflux, ATP-depleted
cells in 2 ml of 50 mM KPi buffer containing 5 mM MgSO4 at
pH 7.0 were pelleted by centrifugation, and loaded with ethidium
bromide at required concentrations ranging from 0.25 mM to10 mM for 1 h at 30uC in a final volume of 400 ml of the samebuffer. After loading with ethidium, cells were pelleted at 15,000 g
for 2 min. The assay was started by resuspension of the cell pellet
in 2 ml of the same buffer without ethidium bromide, and
fluorescence was followed over time. The efflux rates in Fig. 6D
were calculated over the first 20 s of efflux during which efflux was
linear. Rates of passive efflux by nonexpressing control cells were
subtracted from those obtained for cells expressing Wt LmrP or
LmrP mutants.
Active ethidium efflux was monitored in cells preloaded with
2 mM ethidium bromide [32]. The subsequent addition of 20 mMglucose to the ATP-depleted cells allowed the generation of a Dp(inside alkaline and negative) via H+ extrusion by the FOF1ATPase. The Dp-dependent LmrP-mediated efflux of ethidiumwas then monitored by fluorimetry. To manipulate the compo-
sition of the Dp, the ionophores nigericin and valinomycin wereused at final concentrations of 1 mM each, and were added 3 minbefore the addition of glucose.
Hoechst 33342 TransportTransport of Hoechst 33342 (29-(4-Ethoxyphenyl)-5-(4-methyl-
1-piperazinyl)-2,59-bi-1H-benzimidazole) in ISOVs was conductedas described previously [47]. ISOVs (400 mg of total protein) werediluted into 2 ml of 100 mM KPi buffer (pH 7.0) containing
5 mM MgSO4, 0.1 mg/ml creatine kinase, and 5 mM phospho-
creatine (both Roche Diagnostics Ltd, Burgess Hill, UK). After
50 s, Hoechst 33342 (0.25 mM) was added, and the fluorescencewas monitored until a steady state was reached. Subsequently,
sodium ATP (in 50 mM KPi (pH 7.0)) was added to a final
concentration of 2 mM, and the fluorescence intensity was
followed further over time. Hoechst 33342 fluorescence was
measured at excitation and emission wavelengths of 355 and
457 nm, respectively, and slit widths of 10 and 5 nm, respectively.
Measurement of Intracellular pH (pHin) during FacilitatedBenzalkonium Influx
The ATP-depleted cells were prepared as described under
‘‘Ethidium transport in intact cells’’. Subsequently, the cells were
loaded with the fluorescent pH-sensitive probe, 5 (and 6-)-
carboxyfluorescein diacetate N-succinimidyl ester, CFDASE, asdescribed previously [48–50] with modifications. The cells were
incubated for 30 min at 30uC in the presence of 30 mM CFDASE,washed, and resuspended in 50 mM KPi buffer (pH 7.0)
containing 5 mM MgSO4. To eliminate unconjugated CFSE,
10 mM glucose was added, and the cells were incubated for an
additional 30 min at 30uC. The cells were then washed twice,resuspended in the 50 mM KPi buffer (pH 7.0) containing 5 mM
MgSO4, concentrated to a final A660 value of 20, and placed on ice
until required.
To measure the effect of benzalkonium chloride on the pHin,
the concentrated CFSE-loaded cell suspension was diluted 40-fold
in 2 ml of 50 mM KPi buffer (pH 7.0) containing 5 mM MgSO4.
Fluorescence was measured at excitation and emission wave-
lengths of 490 nm 520 nm respectively and slit widths of 5 nm
each. After 200 s, 15 mg/ml benzalkonium chloride (correspond-ing to 44.1 mM) was added. The assay mixture in the cuvette wasthen immediately used for calibration of the fluorescence. After the
Dy and DpH were dissipated in the presence of 1 mM valinomycinplus 1 mM nigericin (pHin = pHout), respectively, the pH of thebuffer was adjusted with either HCl or NaOH to pH values
between 5.5 and 8.0 using a pH micro-electrode, after which
CFSE fluorescence was measured.
Intramolecular Cysteine CrosslinkingFor each crosslinking reaction, RSOVs were diluted to 2 mg
protein/ml in 100 mM KPi buffer, pH 7.0, containing 5 mMMgSO4 in a reaction volume of 30 ml in microcentrifuge tubeswith pierced lids. To reduce background signals due to crosslinks
formed before benzalkonium chloride addition, 0.5 mM DTT was
added, and the samples were incubated for 2 min at 10uC [51].Benzalkonium chloride was then added to a final concentration of
33.3 mg/ml (corresponding with 97.9 mM), and the crosslinkingreactions were initiated by the addition of 0.5 mM copper
phenanthroline solution, which was added from a 50 mM stock
made by mixing CuSO4 and 1,10 phenanthroline in a ratio of 1:4
(w:w) in ultrapure H2O. Some minor precipitates during the
crosslinking reactions were re-dissolved by mixing gently, and the
reaction was allowed to occur at 10uC at incubation times of 5 min(Fig. 9B, D) or 10 min (Fig. 9C). The reactions were stopped by
the addition of excess (10 mM) of the thiol alkylator N-
ethylmaleimide and incubation for 2 min at 10uC. To preventcleavage of formed disulfide crosslinks, 5 mg of protein of RSOVsfrom the crosslinking reactions was mixed with 5x SDS sample-
loading buffer devoid of DTT, and separated on an 11% SDS-
PAGE without any incubation. Gels were analysed by Western
Blotting using anti-His5-tag antibody as described under ‘‘Immu-
noblotting’’. Band intensities on Western blot were compared by
densitometric analyses using ImageJ software version 1.43 (NIH).
Statistical MethodsThe kinetics of facilitated ethidium influx and efflux was fitted to
a hyperbola using the equation: y = ax/(b + x), in which y is thetransport rate in a.u./s, x is the ethidium concentration in mM, a isthe maximum rate of the reaction (Vmax) in a.u./s, and b
(Michaelis constant, Km) is the ethidium concentration (mM) givingKVmax. The calibration curves that provide the relationship
between the pHin and the CFSE fluorescence in cells were fitted
according to a four parameter sigmoid function as described
previously [48]. All experiments were performed at least three
times on independent occasions. Significance was calculated using
the student t-test and ANOVA. Statistical significance was
calculated at a confidence interval of at least 95%.
Acknowledgments
We thank current and previous members of the lab for technical support
and discussions, and Dr Lucy Crouch for critical reading of the manuscript.
Author Contributions
Conceived and designed the experiments: WW HWvV. Performed the
experiments: WW. Analyzed the data: WW HWvV. Contributed reagents/
materials/analysis tools: HWvV. Wrote the paper: WW HWvV.
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 11 June 2012 | Volume 7 | Issue 6 | e38715
References
1. Piddock LJ (2006) Clinically relevant chromosomally encoded multidrug
resistance efflux pumps in bacteria. Clin Microbiol Rev 19: 382–402.2. Nikaido H, Pages JM (2012) Broad-specificity efflux pumps and their role in
multidrug resistance of Gram-negative bacteria. FEMS Microbiol Rev 36: 340–363.
3. Yasufuku T, Shigemura K, Shirakawa T, Matsumoto M, Nakano Y, et al. (2011)
Correlation of overexpression of efflux pump genes with antibiotic resistance inEscherichia coli strains clinically isolated from urinary tract infection patients. J ClinMicrobiol 49: 189–194.
4. Gottesman MM, Fojo T, Bates SE (2002) Multidrug resistance in cancer: role of
ATP-dependent transporters. Nat Rev Cancer 2: 48–58.
5. Paulsen IT, Brown MH, Skurray RA (1996) Proton-dependent multidrug effluxsystems. Microbiol Rev 60: 575–608.
6. Borges-Walmsley MI, McKeegan KS, Walmsley AR (2003) Structure andfunction of efflux pumps that confer resistance to drugs. Biochem J 376: 313–
338.7. Poole K (2005) Efflux-mediated antimicrobial resistance. J Antimicrob Che-
mother 56: 20–51.
8. van Veen HW, Konings WN (1997) Multidrug transporters from bacteria toman: similarities in structure and function. Semin Cancer Biol 8: 183–191.
9. Saier MH Jr, Paulsen IT (2001) Phylogeny of multidrug transporters. Semin CellDev Biol 12: 205–213.
10. Bolhuis H, Poelarends G, van Veen HW, Poolman B, Driessen AJ, et al. (1995)
The lactococcal lmrP gene encodes a proton motive force-dependent drugtransporter. J Biol Chem 270: 26092–26098.
11. Bolhuis H, van Veen HW, Brands JR, Putman M, Poolman B, et al. (1996)Energetics and mechanism of drug transport mediated by the lactococcal
multidrug transporter LmrP. J Biol Chem 271: 24123–24128.12. Bapna A, Federici L, Venter H, Velamakanni S, Luisi B, et al. (2007) Two
proton translocation pathways in a secondary active multidrug transporter. J Mol
Microbiol Biotechnol 12: 197–209.13. Mazurkiewicz P, Poelarends GJ, Driessen AJ, Konings WN (2004) Facilitated
drug influx by an energy-uncoupled secondary multidrug transporter. J BiolChem 279: 103–108.
14. Guan L, Kaback HR (2006) Lessons from lactose permease. Annu Rev Biophys
Biomol Struct 35: 67–91.15. Yamashita A, Singh SK, Kawate T, Jin Y, Gouaux E (2005) Crystal structure of
a bacterial homologue of Na+/Cl–dependent neurotransmitter transporters.Nature 437: 215–223.
16. Shimamura T, Weyand S, Beckstein O, Rutherford NG, Hadden JM, et al.(2010) Molecular basis of alternating access membrane transport by the sodium-
hydantoin transporter Mhp1. Science 328: 470–473.
17. Krishnamurthy H, Gouaux E (2012) X-ray structures of LeuT in substrate-freeoutward-open and apo inward-open states. Nature 481: 469–474.
18. Schuldiner S (2009) EmrE, a model for studying evolution and mechanism ofion-coupled transporters. Biochim Biophys Acta 1794: 748–762.
19. Sigal N, Fluman N, Siemion S, Bibi E (2009) The secondary multidrug/proton
antiporter MdfA tolerates displacements of an essential negatively charged sidechain. J Biol Chem 284: 6966–6971.
20. Brown MH, Skurray RA (2001) Staphylococcal multidrug efflux protein QacA.J Mol Microbiol Biotechnol 3: 163–170.
21. Huang Y, Lemieux MJ, Song J, Auer M, Wang DN (2003) Structure andmechanism of the glycerol-3-phosphate transporter from Escherichia coli. Science301: 616–620.
22. Schaedler TA, van Veen HW (2010) A flexible cation binding site in themultidrug major facilitator superfamily transporter LmrP is associated with
variable proton coupling. FASEB J 24: 3653–3661.23. Venter H, Velamakanni S, Balakrishnan L, van Veen HW (2008) On the
energy-dependence of Hoechst 33342 transport by the ABC transporter LmrA.
Biochem Pharmacol 75: 866–874.24. de Ruyter PG, Kuipers OP, de Vos WM (1996) Controlled gene expression
systems for Lactococcus lactis with the food-grade inducer nisin. Appl EnvironMicrobiol 62: 3662–3667.
25. von Heijne G (1989) Control of topology and mode of assembly of a polytopic
membrane protein by positively charged residues. Nature 341: 456–458.26. Dowhan W, Bogdanov M (2009) Lipid-dependent membrane protein topogen-
esis. Annu Rev Biochem 78: 515–540.27. Venter H, Shilling RA, Velamakanni S, Balakrishnan L, Van Veen HW (2003)
An ABC transporter with a secondary-active multidrug translocator domain.Nature 426: 866–870.
28. Putman M, Koole LA, van Veen HW, Konings WN (1999) The secondary
multidrug transporter LmrP contains multiple drug interaction sites. Biochem-
istry 38: 13900–13905.
29. Steinfels E, Orelle C, Fantino JR, Dalmas O, Rigaud JL, et al. (2004)
Characterization of YvcC (BmrA), a multidrug ABC transporter constitutively
expressed in Bacillus subtilis. Biochemistry 43: 7491–7502.
30. Yu JL, Grinius L, Hooper DC (2002) NorA functions as a multidrug efflux
protein in both cytoplasmic membrane vesicles and reconstituted proteolipo-
somes. J Bacteriol 184: 1370–1377.
31. Shapiro AB, Ling V (1995) Reconstitution of drug transport by purified P-
glycoprotein. J Biol Chem 270: 16167–16175.
32. Balakrishnan L, Venter H, Shilling RA, van Veen HW (2004) Reversible
transport by the ATP-binding cassette multidrug export pump LmrA: ATP
synthesis at the expense of downhill ethidium uptake. J Biol Chem 279: 11273–
11280.
33. Shilling R, Federici L, Walas F, Venter H, Velamakanni S, et al. (2005) A critical
role of a carboxylate in proton conduction by the ATP-binding cassette
multidrug transporter LmrA. FASEB J 19: 1698–1700.
34. Woebking B, Reuter G, Shilling RA, Velamakanni S, Shahi S, et al. (2005)
Drug-lipid A interactions on the Escherichia coli ABC transporter MsbA.
J Bacteriol 187: 6363–6369.
35. Zhou Y, Guan L, Freites JA, Kaback HR (2008) Opening and closing of the
periplasmic gate in lactose permease. Proc Natl Acad Sci U S A 105: 3774–3778.
36. Zhou Y, Nie Y, Kaback HR (2009) Residues gating the periplasmic pathway of
LacY. J Mol Biol 394: 219–225.
37. Abramson J, Smirnova I, Kasho V, Verner G, Kaback HR, et al. (2003)
Structure and mechanism of the lactose permease of Escherichia coli. Science 301:610–615.
38. Shinzawa-Itoh K, Aoyama H, Muramoto K, Terada H, Kurauchi T, et al.
(2007) Structures and physiological roles of 13 integral lipids of bovine heart
cytochrome c oxidase. EMBO J 26: 1713–1725.
39. Lange C, Nett JH, Trumpower BL, Hunte C (2001) Specific roles of protein-
phospholipid interactions in the yeast cytochrome bc1 complex structure.
EMBO J 20: 6591–6600.
40. Schmidt D, Cross SR, MacKinnon R (2009) A gating model for the archeal
voltage-dependent K(+) channel KvAP in DPhPC and POPE:POPG decanelipid bilayers. J Mol Biol 390: 902–912.
41. Schmidt D, Jiang QX, MacKinnon R (2006) Phospholipids and the origin of
cationic gating charges in voltage sensors. Nature 444: 775–779.
42. Long SB, Tao X, Campbell EB, MacKinnon R (2007) Atomic structure of a
voltage-dependent K+ channel in a lipid membrane-like environment. Nature450: 376–382.
43. Brohawn SG, del Marmol J, MacKinnon R (2012) Crystal structure of the
human K2P TRAAK, a lipid- and mechano-sensitive K+ ion channel. Science335: 436–441.
44. Gbaguidi B, Mazurkiewicz P, Konings WN, Driessen AJ, Ruysschaert JM, et al.
(2004) Proton motive force mediates a reorientation of the cytosolic domains of
the multidrug transporter LmrP. Cell Mol Life Sci 61: 2646–2657.
45. Hakizimana P, Masureel M, Gbaguidi B, Ruysschaert JM, Govaerts C (2008)
Interactions between phosphatidylethanolamine headgroup and LmrP, a
multidrug transporter: a conserved mechanism for proton gradient sensing?
J Biol Chem 283: 9369–9376.
46. Kaback HR (1971) Bacterial membranes. Methods Enzymol 22: 99–120.
47. Woebking B, Velamakanni S, Federici L, Seeger MA, Murakami S, et al. (2008)
Functional role of transmembrane helix 6 in drug binding and transport by the
ABC transporter MsbA. Biochemistry 47: 10904–10914.
48. Breeuwer P, Drocourt J, Rombouts FM, Abee T (1996) A novel method for
continuous determination of the intracellular pH in bacteria with the internally
conjugated fluorescent probe 5 (and 6-)-carboxyfluorescein succinimidyl ester.
Appl Environ Microbiol 62: 178–183.
49. Yokota A, Veenstra M, Kurdi P, van Veen HW, Konings WN (2000) Cholate
resistance in Lactococcus lactis is mediated by an ATP-dependent multispecificorganic anion transporter. J Bacteriol 182: 5196–5201.
50. Velamakanni S, Lau CH, Gutmann DA, Venter H, Barrera NP, et al. (2009) A
multidrug ABC transporter with a taste for salt. PLoS One 4: e6137.
51. Doshi R, Woebking B, van Veen HW (2010) Dissection of the conformational
cycle of the multidrug/lipidA ABC exporter MsbA. Proteins 78: 2867–2872.
Role of Basic Residues in Drug Transport by LmrP
PLoS ONE | www.plosone.org 12 June 2012 | Volume 7 | Issue 6 | e38715